Serum-derived EVs (including exosomes and microvesicles) from GBM patients afford an opportunity to longitudinally study tumor evolution and (non)response to therapies in realtime. Currently available diagnostic methods are all based on ?bulk measurements? requiring 103-104 EVs per biomarker for protein measurements (e.g. Western, ELISA) and 102-103 EVs for the more sensitive methods developed by our group (NMR, nPLEX (Nat Biotechnol 2014;32:490-5; Nat Med 2012;18:1835-40). Measurements are thus invariably ?contaminated? and currently do not allow precise analyses of single vesicles. Such analyses however, could be extremely valuable to study EV biogenesis, tumor heterogeneity, rare tumor subtypes, phenotypic changes occurring during therapy, and deeper analyses of host EVs that occur concomitantly with tumoral changes. We have recently developed a single EV analysis (SEA) technique that is capable of measuring multiple protein biomarkers in individual EVs. The goal of this project is to further advance the single EV analysis platform for profiling human GBM-derived EVs from 3D PDX neurosphere models, isogenic murine GBM models, and ultimately in human biofluid samples. We propose three specific aims: i) single EV analysis in neurosphere subtype models; ii) single EV analysis in syngeneic mouse models to allow unbiased profiling of both tumoral and host EVs and iii) clinical analysis in three subsets of patients: a) freshly diagnosed GBM prior to any treatment, b) from glioma patients undergoing standard-of-care (TMZ/avastin/radiation) and c) in patient from the Ad-tk/valacyclovir anti- PD-1 immunotrial in Project 3. In addition to serving as a biomarker discovery platform, analysis of different protein content of EVs will provide insight into how many different types of EVs there are based on biogenesis, whether the tumor-derived EV are unique from the normal cell-derived ones, how they change the phenotype of other normal and tumor cells in support of tumor growth, and possibly how they carry RNA. This project is highly interactive with the Breakefield and Chiocca groups (Project 1 & 3) in identifying protein biomarkers of interest and in evaluating the effects of new EV therapeutics (Cores B & C).